Bitcoin Forum

I have been mining with my Antminer S2's at 200M /0725 Volt with the latest version of the factory firmware. To maintain profitability of the S2's I switched to Time Of Use electricity rate... which about 8 months ago, it was $0.05 KW/h.

The power company did some rate plan editing and consolidated the 3 tiers of TOU into a one size fits all plan... of course, that increased the electric rate up to $0.0864 KW/h. Now my Antminer S2's are only breaking even. I have wood heat which is free... I don't need electric heaters.

I just updated the firmware to Kano's latest firmware and I notice some underclocking options, even a -half speed- option.

I'm assuming that there MIGHT be a little bit of an efficiency curve since I notice that the hotter they run, the amperage draw goes up sharply.... if I underclock, the heat produced should go down, as well as the hashrate, however, since the high heat causes inefficiency what does the efficiency look like when its running cooler.  I'm wondering is anyone has run the underclocking numbers -- clock speed vs hashrate vs amperage( or wattage )

These are some durable miners in my opinion. I have abused these in a couple office rentals with bad ventilation ( but free electricity ) they just do not want to die... If I can squeeze out some better efficiency and make them again profitable to run, even if its $30 in profit a month, I'll run these till they kroak or it becomes permanently unprofitable. Shoot, $30 / mo is BEER money!

Any efficiency tips would be great..... The cheap offices with free electricity have kinda dried up around here.

Thanks

The S2 you can't undervolt. You can underclock but your power efficiency rate is fixed at 1W/1GHS.

You can have it run at full speed at 1Th/s for 1000W or 500Gh/s for 500W.

Underclocking, as a general rule, has almost no effect on efficiency - might help a hair if you can downspeed your fans enough, might get a bit worse if you can't, and the controller will tend to use a (small) fixed amount of power.

 UnderVOLTING is where efficiency gains can be made - but you often have to underclock to get an undervolted unit to run at all.

 If you're on a ToD rate, you might see if you can set up a CRON job in your miner controller to turn off mining (and fans etc) when the rate goes high, and back on when the rate goes low.

I have the whole 240VAC power system setup with a PLC and Allen Bradley controls... Yeah, I know severe overkill using industrial controls, but, its all what I had laying around and is networked so I can remotely through the internet turn on or off each miner or see what amperage each miner is using, it has individual current transformers for individual miner amperage reading, the miners turn on with a 5 second delay between each miner power up to keep thing inrush current at a minimum and less chance for a miner throwing a fit because of a brownout situation when ALL of them turn on at once and the power system runs on a timer using internet (government time server) timing.

I don't run the S2's during peak time, they automatically turn on when the cheap rate is active and automatically turns off when the peak rate starts.

I was hoping that maybe, just maybe, I could squeeze some more $$$ out of these by running these at a lower clockspeed or undervolted... but my quickie tests show a pretty symmetrical curve of clockspeed / wattage / hashrate.

 I'll have to wait till winter to try using Mother Nature for some super douper mega cooling to see if having 20 degree F air running across the chips has any effect ---- putting the miners outside in a weather tight enclosure with direct venting of the intake and exhaust to the great outdoors. ( My contribution towards global warming :-P )

And, yeah, I know .... I need to get newer miners. I just never seem to have money I can blow at new miners when they come out. However, I'm going to try and get a few S7's in the next couple days...

The S2 can be undervolted. I'll have to dig it up but I did a full analysis of the S1 with machine-level efficiency curves from stock all the way down to about 0.6V and there's still some stable operation below the S2's setpoints. The power circuitry is pretty much identical between the two, with the exception of volt-setting resistors, but proper values can easily be calculated to get another 10-15% of efficiency (possibly, depends on chip quality) out of the S2.

IIRC the S2 has to be undervolted by modding the hardware, pretty much the same as the S1 on that mod.